Advantages of the KDS/BCA Assay over the Bradford Assay for Protein Quantification in White Wine and Grape Juice

Transcription

1 Advantages of the KDS/BCA Assay over the Bradford Assay for Protein Quantification in White Wine and Grape Juice Diana Gazzola, 1 Simone Vincenzi, 1 * Gabriella Pasini, 1 Giovanna Lomolino, 1 and Andrea Curioni 1 Abstract: The present study compared the performance of two colorimetric protein assays, the Bradford and the potassium dodecyl sulfate/bicinchoninic acid (KDS/BCA) assays, for use in wine and grape juice analysis. The Bradford assay was affected by protein type, whereas the KDS/BCA assay had lower protein-to-protein variation. Bovine serum albumin and lysozyme yielded an absorbance (562 nm) vs. protein concentration slope (dose response curve) similar to that of wine proteins. In the Bradford assay, the presence of 12% ethanol and 200 mg/l of wine polyphenols decreased the protein absorbance by 28 and 16%, respectively, whereas in the KDS/BCA assay such interference was not significant. Among 64 white wines, the correlation between protein haze potential, determined by a heat test, and protein content was better for the KDS/BCA assay. This study confirmed the superiority of the KDS/BCA assay over the Bradford assay for quantifying protein in white grape juice and wine, and it yielded better predictive value with respect to the risk of white wine protein instability. Key words: colorimetric assays, protein quantification, white wine, wine haze Grape proteins play a central role in winemaking. Although most grape proteins are presumed to precipitate during fermentation or postfermentation storage, many remain soluble or transiently soluble in bottled wine (Waters et al. 2005). Winemakers are most interested in removing the proteins responsible for haze formation in white wines, typically by bentonite fining (Waters et al. 2005); however, these proteins in smaller concentrations can make desirable contributions to wine quality, e.g., foaming properties of sparkling wines (Brissonnet and Maujean 1991) and interactions with aroma compounds (Lubbers et al. 1994). Several colorimetric assays are available for quantifying protein content; however, three drawbacks confound their use with grape juice and wine (Le Bourse et al. 2010). Proteins are typically present at low concentrations in these media ( mg/l) (Ferreira et al. 2001, Vincenzi et al. 2005), and assays with sufficient sensitivity are needed. The presence of potential interfering compounds such as polyphenols and ethanol limit the choice of protein assay (Compton and 1 Department of Agronomy, Food, Natural Resources, Animals and Environment (DAFNAE) and Centro Interdipartimentale per la Ricerca in Viticoltura ed Enologia (CIRVE), Università di Padova, via dell Università 16, Legnaro (PD), Italy. *Corresponding author (simone.vincenzi@unipd.it; tel: (+39 0) ; fax (+39 0) ) Acknowledgment: The authors would like to thank Alan T. Bakalinsky (Oregon State University) for his assistance in writing this paper. Manuscript submitted Jul 2014, revised Oct 2014, accepted Nov 2014 Publication costs of this article defrayed in part by page fees. Copyright 2015 by the American Society for Enology and Viticulture. All rights reserved. doi: /ajev Jones 1985, Marchal et al. 1997). Further, the protein standard chosen for a given assay may not be representative of the proteins found in grape juice or wine (Le Bourse et al. 2010). Among the assays available to quantify protein in grape juice and wine, the Bradford assay (Bradford 1976), based on Coomassie Brilliant Blue G-250 (CBB), is probably the most widely used assay (Brissonnet and Maujean 1991, Waters et al. 1991, Boyes et al. 1997). Its popularity is primarily due to its simplicity and rapidity (Marchal et al. 1997). It requires the addition of a single reagent to the sample and a relatively short incubation time before the absorbance of the sample is measured. The assay is based on the immediate absorbance shift from 470 nm to 595 nm that occurs when the CBB dye binds to protein under acidic conditions. This binding is believed to occur via Van der Waals and hydrophobic interactions (Compton and Jones 1985). Notably, the Bradford assay has limitations and sometimes results in overestimates or underestimates of actual protein content (Compton and Jones 1985). Overestimation occurs when nonprotein molecules stabilize the neutral species of the dye (A max 650 nm), which contributes to absorbance at 595 nm. Compounds associated with this phenomenon include detergents such as sodium dodecyl sulfate (SDS) and phenolics extracted from grape skin (Marchal et al. 1997). The presence of these compounds shifts the equilibria among the three species (red cationic species, green neutral species, and blue anionic species). Known compounds that overestimate protein detection include those that stabilize the green neutral species (A max 650 nm) by direct binding to the dye. By contrast, other compounds can contribute to the underestimation of the protein content. For example, polysaccharides can interfere with the interaction of the dye and the proteins by steric hindrance (Fountoulakis et al. 1992). Unlike 227

2 228 Gazzola et al. the Bradford assay which is applied directly to wine, other colorimetric assays such as the Lowry (Lowry et al. 1951), Biuret (Gornall et al. 1949), or Bicinchoninic acid (BCA, Smith et al. 1985) assays require prior separation of wine proteins from the matrix. This is necessary because of the interference caused by common grape and wine constituents such as glutathione, phenolic compounds, and potassium ions (Godshall 1983, Brenna and De Vecchi 1990). The BCA assay has many advantages over the alternatives, including sensitivity, simplicity, stability of the chromophore, low proteinto-protein variation (Schoel et al. 1995), and insensitivity to many contaminating substances such as commonly used detergents (Smith et al. 1985). Vincenzi et al. (2005) developed a new procedure for the precipitation and the quantification of proteins in grape juice and wine. Different combinations of protein recovery and quantitation methods were tested using a BSA protein standard. The best coupling was found to be between the KDS method for protein recovery and the Smith assay (BCA) for quantification. The resultant assay (KDS/BCA) involved the complexation of proteins with the detergent dodecyl sulfate (DS), the addition of potassium ions (KCl) to decrease the solubility of the protein-ds complexes, and the colorimetric quantification of the proteins, as potassium dodecyl sulfate (KDS) complexes, with the BCA reagent. This specific sequence of steps proved to be the most accurate among the different assays tested, including the Bradford assay (Vincenzi et al. 2005). In the present study, a more comprehensive comparison of the KDS/BCA assay with the Bradford assay to detect wine proteins was undertaken. Differences from the previous work of Vincenzi et al. (2005) were that ethanol and polyphenol compounds were included as potential interfering substances, and the choice of an appropriate protein standard was addressed. In addition, the protein content of several white wines was calculated using both assays to assess which of them better predicted wine protein haze potential. Materials and Methods The model wine (MW) contained 5 g/l tartaric acid (Baker, Deventer, Netherlands) and 12% ethanol (VWR Prolabo, Fontenay Sous Bois, France) buffered to ph 3.20 with NaOH (Carlo Erba, Milan, Italy). Unfined Sauvignon blanc (10.56% v/v ethanol) and Manzoni bianco (12.20% v/v ethanol) wines and Glera grape juice were supplied by Scuola Enologica G.B. Cerletti, Conegliano (Italy). Wines and juice were not subjected to bentonite fining. They were ultrafiltered using an Amicon RC800 apparatus with a regenerated cellulose membrane (3 kda MWCO; Millipore Corporation, Bedford, MA) and stored at 4 C until use. The 64 white wines included in the study were obtained from different wineries located in the Veneto region (Italy) and were a mix of both international and local varieties. All wines were prepared using standard winemaking practices on a commercial scale and were collected before bentonite fining. The standard proteins bovine serum albumin (BSA), Thaumatin from Thaumatococcus danielli (THAU), and ovalbumin (OVA) were purchased from Sigma-Aldrich (Steinheim, Germany). Lysozyme (LYS) was purified from an enological lysozyme preparation (Oliver Ogar Italia, Verona, Italy) using an S-Sepharose Fast Flow column (Pharmacia Biotech, Uppsala, Sweden). LYS purity was assessed by RP6 HPLC (Waters 1525 HPLC system; Waters, Milford, MA) on a Vydac C18 column (Anaheim, CA) according to Van Sluyter et al. (2009). Manzoni bianco wine was used as the starting material for isolation of wine polyphenols and wine total proteins. Isolation of wine polyphenols. Polyphenols were extracted from multiple 10-mL aliquots of wine that were initially filtered (0.2 µm cellulose acetate filters; Sartorius, Gottingen, Germany), and subsequently ultrafiltered (3 kda MWCO). A Sep-Pak C18 cartridge (10 g; Waters, Milford) was prewashed with 50 ml of methanol (Carlo Erba) and equilibrated with 50 ml of deionized water. The ultrafiltered wine was then loaded onto the cartridge. The unbound fraction was removed with 100 ml of water, and polyphenols were eluted in 30 ml of methanol. The solvent was removed in a rotary evaporator (BuchiRotavapor R-114, Flawil, Switzerland) at 35 C. The residue was resuspended in 10 ml of MW and the concentration of polyphenols was estimated by the Folin-Ciocalteu assay (Singleton and Rossi 1965). Purification of wine proteins. Wine total proteins (WTP) were recovered from 5 L of unfined Manzoni bianco wine as follows. The wine was treated at 4 C with 4 g/l of polyvinylpolypyrrolidone (Carlo Erba). After 24 hr at 15 C, the wine was filtered through a GF/A glass microfiber filter (Whatman, Maidstone, UK), adjusted to ph 3.0 with HCl (Carlo Erba) and then filtered through a 0.2 μm cellulose acetate membrane (Sartorius). WTP were purified by strong cation exchange chromatography as described by Van Sluyter et al. (2009) with the following modifications: chromatography was performed on a S-Sepharose Fast Flow (Pharmacia) column (5 x 14 cm), equilibrated with 30 mm sodium citrate (Carlo Erba), ph 3.0 (buffer A), at a flow rate of 8 ml/min. Five liters of wine were loaded on the column, which was then washed with buffer A. Proteins were eluted using a gradient consisting of 30 mm MES (Sigma-Aldrich, St. Louis, MO), 1 M NaCl, ph 6.0 (buffer B) (0 to 40% buffer B in 120 min, then 100% buffer B). All protein fractions were pooled and concentrated using a stirred ultrafiltration cell (Amicon RC800 apparatus) equipped with 3 kda membrane (Millipore Corporation). The retentate, containing the WTP, was dialyzed against water in 3.5 kda MWCO regenerated cellulose membrane bags (CelluSep, Membrane Filtration Products Inc., Seguin, TX). Protein standard curves. Stock solutions of BSA, LYS, THAU and WTP (at 10 mg/ml), and OVA (at 5 mg/ml) were prepared in distilled water and stored at 4 C. Standard solutions of each protein were prepared: 2.5, 5.0, 10.0, 15.0, 20.0, 30.0, 40.0, 60.0, and 80.0 µg/ml for the Bradford assay and 62.5, 125.0, 250.0, 500.0, and 1,000.0 µg/ml for the KDS/ BCA assay. These concentrations fall within the optimal working range for each of the assays. Standard curves were freshly prepared individually, and each experiment was repeated five times. Proteins were solubilized in different matrices: water (H 2 O); MW; ultrafiltered Glera grape juice (GJ);

3 Protein Quantification in White Wine 229 and two ultrafiltered (3 kda MWCO) wines, Sauvignon (SAU UW) and Manzoni bianco (MB UW). The mean dose response curve (absorbance vs. concentration) slope obtained for the four standard proteins (BSA, THAU, OVA, and LYS) in water or in each of the four different matrices (MW, ultrafiltered GJ, SAU UW, and MB UW) were evaluated. In a second set of experiments, the standard proteins were diluted in MW containing increasing amounts of total Manzoni bianco polyphenols (0, 25, 50, 100, 200 mg/l) to ascertain the effects of polyphenols on the mean slope of the dose response curve obtained using the standard proteins. The values for the slopes of the dose response curves obtained using the standard proteins or WTP in MB UW were calculated and expressed as A 595 /µg protein (Bradford assay) or A 562 /µg protein (KDS/BCA assay). Five replicates for each protein dissolved in each matrix were performed. The coefficient of linear regression (R 2 ) and the y-intercept were computed. Only values for slopes obtained from calibration curves with goodness-of-fit (R 2 ) values >0.99 were used. Protein quantification using the KDS/BCA assay. Proteins were recovered from samples by KDS precipitation as described (Vincenzi et al. 2005) with minor modifications. This system involved treatment of samples with SDS and the addition of KCl, which precipitates proteins as insoluble KDS complexes. Briefly, 5 µl of a 10% solution (w/v) of SDS (Bio- Rad Laboratories Inc., Hercules, CA) were added to 500 µl of sample. The mixture was heated at 100 C for 5 min after which 125 ml of 1 M KCl (Carlo Erba) were added. After 2 hr of incubation at room temperature, protein pellets were collected by centrifugation (12,000 g, 15 min, 4 C). These were washed three times with 1 ml of 1 M KCl and freeze-dried. Prior to use for protein determination, freeze-dried protein pellets were resuspended in 500 µl of water and heated for 5 min to improve protein solubilization. Protein was quantified using the BCA-200 Protein Assay kit (Pierce, Rockford, IL), which is based on the method described by Smith et al. (1985). The absorbance of samples was measured at 562 nm (Ultrospec 2100 pro UV vis; Amersham Biosciences, Uppsala, Sweden). Protein quantification using the Bradford assay. The Bradford protein assay was performed using a commercial kit (Bio-Rad Laboratories Inc.). Protein samples (400 µl) dissolved in the various matrices were mixed with an equal volume of deionized water to which 200 ml of Bio-Rad Protein assay reagent were added. A 595 readings were measured after holding samples at room temperature for 1 hr (Marchal et al. 1997). Heat test. The protein stability of 64 white wines was determined using a heat test (Pocock and Rankine 1973). White wine samples were heated at 80 C for 6 hr, held at 4 C for 12 hr, and then transferred to room temperature. After 20 min at room temperature, haze was measured with a HI Turbidity and Bentocheck Meter (Hanna Instrument, Szeged, Hungary). The protein stability of each wine was determined by calculating the difference between the turbidity units (NTU) measured before and after the heat test. Three replicates for each wine were analyzed. Statistical analysis. Statistical analysis was performed using CoHort Software (Costat version 6.4, Monterey, CA). Data were evaluated by one-way (data sets in experiments 1 and 4, see below) and two-way (data sets in experiments 2 and 3) completely randomized analysis of variance (ANOVA). Tukey s HSD test was used to compare the means when significant differences were found by ANOVA (p 0.05 was used as criterion for statistical significance). The relationship between the wine protein content vs. the heat test turbidity variables was evaluated by the Pearson test (p 0.05) (experiment 5). Results and Discussion Protein-to-protein variability: comparison of the dose response curves of different proteins dissolved in water (experiment 1). The reactivity of aqueous solutions of increasing concentrations of four standard proteins (BSA, LYS, OVA, and THAU) in both the Bradford and KDS/BCA assays was evaluated by calculating the slope obtained from dose response curves. In the Bradford assay, at any given protein concentration, LYS yielded the highest absorbance, OVA yielded the lowest absorbance, and BSA and THAU yielded intermediate absorbance values, which were similar to each other (Figure 1). Figure 1 Mean response of the slope of the dose response calibration curves obtained for bovine serum albumin (BSA), lysozyme (LYS), ovalbumin (OVA), and Thaumatococcus daniellii thaumatin (THAU) dissolved in water and measured using either the Bradford (A) or the KDS/BCA (B) assay. The main effect was significant at p as determined by ANOVA. Bars with different letters were significantly different as determined by Tukey s HSD test (p 0.05).

4 230 Gazzola et al. The absorbance obtained using CBB (the Bradford assay dye) is protein-dependent because it varies with amino acid composition (Tal et al. 1985, Compton and Jones 1985); lysine, arginine, and histidine play an important role (Noble et al. 2007). LYS contains a large percentage of arginine residues (8.2%) for which CBB has high binding affinity (Compton and Jones 1985). Conversely, the glycoprotein OVA, which contains a small percentage of aromatic (Phe, Trp and Tyr: 8.6%) and basic (Arg, His and Lys: 10.9%) amino acids, binds CBB with low affinity (Antharavally et al. 2009). CBB also results in underestimates of glycoprotein concentrations because of steric hindrance caused by glycans or their effects on protein hydrophobicity (Fountoulakis et al. 1992). THAU, a good model for wine proteins because of its similarity to the grape thaumatin-like proteins (TLPs) (Edens et al. 1982), had a dose response curve slope similar to that of BSA, a common protein standard (Noble et al. 2007). In the KDS/BCA assay, LYS and THAU had similar but higher dose response curve absorbance values than BSA and OVA (with values that were similar to one another) (Figure 1). Color formation in the BCA assay is influenced by the presence of Tyr, Trp, and Cys (Smith et al. 1985). LYS contains 6.1% Cys and 4.1% Trp, and THAU contains 7.7% Cys and 3.9% Tyr; BSA and OVA have lower percentages of those amino acids, which provides a possible explanation for their differences in absorbance. Overall, aqueous solutions of the tested proteins exhibited less protein-to-protein variability in the KDS/BCA assay (CV = 19.27%; calculated by considering the absorbances of all the standard proteins) than in the Bradford assay (CV = 27.12%) (Table 1). Matrix variability: evaluation of matrix effects on protein dose response curves (experiment 2). An important problem in protein quantification of grape juices and wines using colorimetric assays is the presence of compounds that interfere with the protein-dye binding. The effect of five different matrices on protein absorbance was evaluated: water (H 2 O); MW; ultrafiltered Glera grape juice (GJ); and two ultrafiltered (3 kda MWCO) wines, SAU UW and MB UW. Using the Bradford assay, the proteins yielded higher absorbances in MW than in water (data not shown) indicating interference caused by ethanol. Marchal et al. (1997) found that alcohol caused a significant increase in absorbance and suggested using a standard blank solution containing the same amount of alcohol as the samples to be analyzed. However, results herein show that ethanol decreased the slope of the protein dose response curve (28%) when compared to the dose response curve of proteins prepared in water (Figure 2). This result is consistent with the possibility that ethanol increased the absorbance of the dye (Marchal et al. 1997), changed the dye-protein interactions, or both. The presence of ethanol favors the neutral form of the dye at the expense of the ionized form of the dye, which is the form that is reactive with proteins (Marchal et al. 1997). The dose response curve of proteins in GJ yielded the same mean absorbance slopes as those in MW (Figure 2). Because there was no ethanol in GJ, this result could be caused by the presence of other interfering substances such as sugars and/or polyphenols (Compton and Jones 1985, Godshall 1983, Brenna and De Vecchi 1990). The mean slope of the dose response curve for proteins dissolved in two different ultrafiltered white wines (MB UW and SAU UW) was not significantly different from that obtained using samples of proteins in MW (Figure 2). However, the protein absorbance measured in SAU UW was higher than that in MB UW, perhaps because of the higher ethanol content in the MB UW. Because the same Table 1 Protein-to-protein variation with the Bradford and the KDS/BCA methods. The mean slope of the dose response curve (A 595 /μg [Bradford assay] or A 562 /μg [KDS/BCA assay]) obtained for lysozyme (LYS), ovalbumin (OVA), and Thaumatococcus daniellii thaumatin (THAU) was normalized with respect to that obtained for bovine serum albumin (BSA; set as 100). Protein tested Bradford assay KDS/BCA assay BSA LYS OVA THAU Coefficient of variation 27.12% 19.27% Figure 2 Mean response of the slope of the dose response calibration curves for proteins in water (H 2 O); model wine (MW); ultrafiltered grape juice (GJ); and ultrafiltered wines, Manzoni bianco (MB UW) and Sauvignon (SAU UW) using the Bradford (A) or the KDS/BCA (B) assay. In A, the main effect matrix was significant at p as determined by ANOVA. Bars with different letters were significantly different as determined by Tukey s HSD test (p 0.05). In B, the main effect matrix is not significant as determined by ANOVA.

5 Protein Quantification in White Wine 231 amount of protein was added to the two ultrafiltered wines, these results showed that the composition of each individual wine affected protein quantification using the Bradford assay. By contrast, the results obtained using the KDS/BCA assay were not affected by matrix composition (Figure 2). Because no statistically significant differences were noted between protein dose response curves in water or in MW, we concluded that ethanol or acidity may not influence the KDS/BCA method. The separation of proteins from the matrix (KDS precipitation) is particularly important in grape juice and wine samples to eliminate interfering compounds. Residual SDS in the protein pellet is compatible with the BCA reagent (Vincenzi et al. 2005) but not with the Bradford reagent, because SDS interferes with the CBB dye. Moreover, using the KDS/BCA assay, it is possible to quantify the protein content of different white wines using a calibration curve prepared in water. The slope of the dose response curve of proteins prepared in water does not differ from that of proteins prepared from grape juice or wine (Figure 2). Notably, for proteins dissolved in water, the slopes of the dose response curve obtained by the KDS/BCA assay were not different from those obtained by omitting the KDS precipitation step (not shown). Effect of polyphenols on the protein dose response curves (experiment 3). Endogenous wine phenolic compounds have previously been shown to interfere with quantification of wine protein (Compton and Jones 1985, Whiffen et al. 2007, Marchal et al. 1997). For example, Marchal et al. (1997) demonstrated that an aqueous buffered solution of phenolic compounds from Pinot noir and Chardonnay skins yielded an absorbance equivalent to that of 16.3 mg/ml and 16.7 mg/l of BSA, respectively, using the Bradford assay. However, this interference was detected in the absence of protein, which eliminated the possibility of measuring other confounding interactions between proteins and phenolic compounds by colorimetric absorbance. The effect of increasing the concentration of wine phenolic compounds from 0 mg/l to 200 mg/l in the presence of protein was evaluated in MW. This range was consistent with the polyphenol content of typical Italian wines (Simonetti et al. 1997). Using the Bradford assay, a decrease (16%) in the slope of the dose response curve was observed at a polyphenol concentration of 200 mg/l. At lower concentrations, no effect was observed (Figure 3). This indicates the unsuitability of the Bradford assay for quantifying proteins in white wines or juices with concentrations of polyphenols 200 mg/l. By contrast, the quantification of proteins using the KDS/ BCA assay was not affected by the presence of up to 200 of mg/l wine polyphenols (Figure 3), demonstrating that the KDS/BCA assay is not affected by wine phenolic compounds at the concentrations tested. Dose-response curve of wine proteins compared to those of standard proteins (experiment 4). The wine total proteins and the standard proteins in MB UW (WTP, BSA, LYS, OVA, and THAU) were quantified using both assays. Less variability was observed with the KDS/BCA assay (CV = 17%) than with the Bradford assay (CV = 29%) (Figure 4). In the Bradford assay, the absorbance of WTP and OVA was the same, suggesting that OVA would make an appropriate standard protein. By contrast, the widely used BSA, which has been thought to be a good model for proteins in wine (Waters et al. 1991), appeared not to be a standard of choice for quantifying protein in wine when using the Bradford assay (Figure 4). In the KDS/BCA assay, THAU yielded the highest absorbance in the dose response curve and OVA yielded the lowest absorbance. WTP, BSA, and LYS yielded similar intermediate absorbances (Figure 4). Therefore, in the KDS/BCA assay, either BSA or LYS may be a suitable standard for the quantification of wine proteins recovered by KDS precipitation because they yield the same absorbance as WTP. By contrast, thaumatin from T. daniellii is not representative of WTP when measured using the KDS/BCA assay even though WTP typically contains a significant fraction of TLPs (Peng et al. 1997). Correlation of total protein content and wine haze potential (experiment 5). The protein content of 64 unfined white wines from the Veneto region was quantified using the Bradford and the KDS/BCA assays. By the Bradford assay (using OVA as standard), the protein content measured ranged from 1.49 mg/ml to mg/l; whereas, by the KDS/BCA assay (using BSA as standard), the protein content measured Figure 3 Mean response of the slope of the dose response calibration curves obtained in the presence of increasing concentrations of polyphenols in MW using the Bradford (A) or the KDS/BCA (B) assay. In A, the main effect polyphenols was significant at p 0.01 as determined by ANOVA. Bars with different letters were significantly different as determined by Tukey s HSD test (p 0.05). In B, the main effect polyphenols was not significant as determined by ANOVA.

6 232 Gazzola et al. ranged from 1.61 mg/ml to mg/l. These results were consistent with the observation that the Bradford assay systematically underestimates protein content (Waters et al. 1991). Although the protein instability of white wines was not completely explained by the way protein concentrations were measured, this parameter has a significant effect on protein haze formation (Sarmento et al. 2000). To verify if the reported low correlation between protein content and wine instability (Dawes et al. 1994) was related to the unsuitability of the assays used to determine protein content, protein (mg/l) was measured by both the Bradford and KDS/BCA assays and the results were correlated with the results of the heat test (NTU), which is a predictor of wine protein instability (Pocock and Rankine 1973). The correlation between the heat test results and protein content measured by the KDS/ BCA assay was higher (Pearson correlation coefficient r = 0.79; p 0.05) than that for the protein content measured by the Bradford assay (Pearson correlation coefficient r = 0.54; p 0.05). One interpretation of this finding is that the KDS/ BCA assay is biased for detection of wine proteins that tend to contribute to protein instability. Fusi et al. (2010) found that the KDS system did not precipitate glycoproteins present in white wines, perhaps because the glycan interfered with the SDS-mediated precipitation step. Similarly, Smith et al. (2011) found that the KDS/ Figure 4 Mean values of the slope of the dose-response calibration curves obtained for bovine serum albumin (BSA), lysozyme (LYS), ovalbumin (OVA), Thaumatococcus daniellii thaumatin (THAU), and wine total protein (WTP) dissolved in Manzoni Bianco ultrafiltered wine, using the Bradford (A) or the KDS/BCA (B) assay. The main effect was significant at p as determined by ANOVA. Bars with different letters were significantly different as determined by Tukey s HSD test (p 0.05). BCA assay detected <15% of the yeast glycoprotein invertase when added to wine. We speculated that the proteins recovered from the wine matrix as KDS complexes were primarily TLPs and chitinases, which have previously been implicated in wine haze formation (Waters et al. 2005). However, only a small quantity of the grape invertase, a predominant protein in wine (Jégou et al. 2009), was recovered by the KDS precipitation. This was also reported in other studies that examined SDS-PAGE results (Vincenzi et al. 2005, Fusi et al. 2010). The presence of a high quantity of mannose-rich glycans linked to certain wine proteins (Palmisano et al. 2010) such as invertase could explain this occurrence. Invertase, unlike TLPs and chitinases, is considered to be a stable macromolecule in wine (Dufrechou et al. 2013) and is thus not involved in haze formation. Therefore, the KDS precipitation may primarily involve only the unstable proteins, which could explain why the KDS/BCA assay better predicts wine haze. The protein quantity determined with the KDS/BCA assay did not fully correlate with the results of the heat test (r-value not equal to 1). This result could have been caused by the presence of other factors (in addition to total protein) that contribute to haze formation in wines (Sarmento et al. 2000, Waters et al. 2005). However, the results showed that nonprotein factors play a smaller role in wine haze formation than total protein. Conclusion The Bradford assay, the technique based on CBB binding to proteins, was found to be unsuitable for the quantification of the protein content in grape juice and wine, which confirms other results. That assay is sensitive to interfering compounds present in the matrix such as ethanol and polyphenols and is affected by the type of protein quantified. By contrast, when staining with BCA, the proteins separated from the wine matrix as KDS complexes (KDS/BCA assay) proved to be superior for protein quantification in grape juices and wines. This assay was cost-effective when the standard proteins used were LYS or BSA, which both showed a dose response curve similar to that of the total wine proteins. It has been demonstrated that the protein concentration, when estimated by the KDS/BCA method, shows a positive correlation with wine haze formation, likely because unstable proteins are primarily quantified. This result reinforces the idea that protein quantity is the main factor involved in white wine hazing, although not the only one. Therefore, the KDS/BCA assay can be recommended for protein quantification in grape juices and wines in both research and technical settings. Literature Cited Antharavally, B.S., K.A. Mallia, P. Rangaraj, P. Haney, and P.A. Bell Quantitation of proteins using a dye-metal-based colorimetric protein assay. Anal. Biochem. 385: Boyes, S., P. Strübi, and H. Dawes Measurement of protein content in fruit juices, wine and plant extracts in the presence of endogenous organic compounds. LWT- Food Sci. Technol. 30: Bradford, M.M A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:

7 Protein Quantification in White Wine 233 Brenna, O., and S. De Vecchi Evaluation of protein and phenolic content in must and wine. I. Assay of soluble proteins. Ital. J. Food Sci. 2: Brissonnet, F., and A. Maujean Identification of some foam-active compounds in champagne base wines. Am. J. Enol. Vitic. 42: Compton, S.J., and C.G. Jones Mechanism of dye response and interference in the Bradford protein assay. Anal. Biochem. 151: Dawes, H., S. Boyes, J. Keene, and D. Heatherbell Protein instability of wines: Influence of protein isoelectric point. Am. J. Enol. Vitic. 45: Dufrechou, M., A. Vernhet, P. Roblin, F.X. Sauvage, and C. Poncet- Legrand White wine proteins: How does the ph affect their conformation at room temperature? Langmuir 29: Edens, L., L. Heslinga, R. Klok, A.M. Ledeboer, J. Maat, M.Y. Toonen, C. Visser, and C.T. Verrips Cloning of cdna encoding the sweet-tasting plant protein thaumatin and its expression in Escherichia coli. Gene 18:1-12. Ferreira, R.B., M.A. Piçarra-Pereira, S. Monteiro, V.B. Loureiro, A.R. Teixeira The wine proteins. Trends Food Sci. Technol. 12: Fountoulakis, M., J.F. Juranville, and M. Manneberg Comparison of the Coomassie brilliant blue, bicinchoninic acid and Lowry quantitation assays, using non-glycosylated and glycosylated proteins. J. Biochem. Biophys. Methods 24: Fusi, M., F. Mainente, C. Rizzi, G. Zoccatelli, and B. Simonato Wine hazing: A predictive assay based on protein and glycoprotein independent recovery and quantification. Food Control 21: Godshall, M.A Interference of plant polysaccharides and tannin in the Coomassie Blue G-250 test for protein. J. Food Sci. 48: Gornall, A.G., C.J. Bardawill, and M.M. David Determination of serum proteins by means of the biuret reaction. J. Biol. Chem. 177: Jégou, S., A. Conreux, S. Villaume, A. Hovasse, C. Schaeffer, C. Cilindre, A. Van Dorsselaer, and P. Jeandet One step purification of the grape vacuolar invertase. Anal. Chim. Acta 638: Le Bourse, D., S. Jégou, A. Conreux, S. Villaume, and P. Jeandet Review of preparative and analytical procedures for the study of proteins in grape juice and wine. Anal. Chim. Acta 667: Lowry, O.H., N.J. Rosebrough, A.L. Farr, and R.J. Randall Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193: Lubbers, S., A. Voilley, M. Feuillat, and C. Charpentier Influence of mannoproteins from yeast on the aroma intensity of a model wine. LWT-Food Sci. Technol. 27: Marchal, R., V. Seguin, and A. Maujean Quantification of interferences in the direct measurement of proteins in wines from the Champagne region using the Bradford method. Am. J. Enol. Vitic. 48: Noble, J.E., A.E. Knight, A.J. Reason, A. Di Matola, and M.J. Bailey A comparison of protein quantitation assays for biopharmaceutical applications. Mol. Biotechnol. 37: Palmisano, G., D. Antonacci, and M.R. Larsen Glycoproteomic profile in wine: A sweet molecular renaissance. J. Proteome Res. 9: Peng, Z., K.F. Pocock, E.J. Waters, I.L. Francis, and P.J. Williams Taste properties of grape (Vitis vinifera) pathogenesis-related proteins isolated from wine. J. Agric. Food Chem. 45: Pocock, K.F., and B.C. Rankine Heat test for detecting protein instability in wine. Aust. Wine Brew. Spirit Rev. 91: Sarmento, M.R., J.C. Oliveira, M. Slatner, and R.B. Boulton Influence of intrinsic factors on conventional wine protein stability tests. Food Control 11: Schoel, B., M. Welzel, and S.H. Kaufmann Quantification of protein in dilute and complex samples: modification of the bicinchoninic acid assay. J. Biochem. Biophys. Methods 30: Simonetti, P., P. Pietta, and G. Testolin Polyphenol content and total antioxidant potential of selected Italian wines. J. Agric. Food Chem. 45: Singleton, V.L., and J.A. Rossi Colorimetry of total phenolics with phosphomolybdic-phosphotungstic acid reagents. Am. J. Enol. Vitic. 16: Smith, M.R., M.H. Penner, S.E. Bennett, and A.T. Bakalinsky Quantitative colorimetric assay for total protein applied to the red wine Pinot noir. J. Agric. Food Chem. 59: Smith, P.K., et al Measurement of protein using bicinchoninic acid. Anal. Biochem. 150: Tal, M., A. Silberstein, and E. Nusser Why does Coomassie Brilliant Blue R interact differently with different proteins? A partial answer. J. Biol. Chem. 260: Van Sluyter, S.C., M. Marangon, S.D. Stranks, K.A. Neilson, Y. Hayasaka, P.A. Haynes, R.I. Menz, and E.J. Waters Two-step purification of pathogenesis-related proteins from grape juice and crystallization of thaumatin-like proteins. J. Agric. Food Chem. 57: Vincenzi, S., S. Mosconi, G. Zoccatelli, C. Dalla Pellegrina, G. Veneri, R. Chignola, A. Peruffo, A. Curioni, and C. Rizzi Development of a new procedure for protein recovery and quantification in wine. Am. J. Enol. Vitic. 56: Waters, E.J., W. Wallace, and P.J. Williams Heat haze characteristics of fractionated wine proteins. Am. J. Enol. Vitic. 42: Waters, E.J., G. Alexander, R. Muhlack, K.F. Pocock, C. Colby, B.K. O Neill, P.B. Hoj, and P. Jones Preventing protein haze in bottled white wine. Aust. J. Grape Wine Res. 11: Whiffen, L.K., D.J. Midgley, and P.A. McGee Polyphenolic compounds interfere with quantification of protein in soil extracts using the Bradford method. Soil Biol. Biochem. 39: